POPULATION-DYNAMICS OF SOYBEAN ROOT-NODULE BACTERIA IN LATOSOL SOIL USED FOR UPLAND AND LOWLAND RICE SOYBEAN CROPPING SYSTEMS IN WEST JAVA,INDONESIA/

Two experiments were established in a latosol soil near Bogor, Indones ia to examine the population dynamics of soybean rhizobia under soybea n-upland and -lowland rice management systems. Rice was sown in all pl ots before sowing the first soybean crop which was inoculated with eit her the wild-type or antibiotic-resistant mutants of strains CB1809, U SDA 110 and LRj I1D or left uninoculated. Numbers of soybean rhizobia in the soil before sowing the rice and in rice rhizospheres after 21 a nd 42 days, before sowing soybeans and in the soybean rhizospheres aft er 28 and 42 days were estimated by MPN using Glycine soja. In the low land treatment, numbers of soybean rhizobia in the bulk soil were low before sowing rice (log(10) 1.86 g(-1) soil) but they multiplied in th e rhizospheres to ca log(10) 4.41 g(-1) soil. In the upland treatment there were log(10) 4.45 initially which increased to log(10) 5.15 at 4 2 days. Following 2 months fallow and soil preparation for sowing soyb eans in the upland treatment, the numbers of soybean rhizobia in the b ulk soil were only log(10) 1.29 g(-1) but they multiplied to log(10) 4 .84 g-l soil at 28 days in the soybean rhizospheres. Numbers in the so ybean rhizospheres in the lowland treatment reached log(10) 5.27 g(-1) at 28 days. They again declined rapidly, after the soybeans were harv ested, to log(10) 1.32 and 1.98 g(-1) of bulk soil in upland and lowla nd systems respectively. Inoculation did not affect nodulation, shoot dry weight or grain yield of soybean despite nodule occupancy by selec ted strains, of between 52-63% in the lowland and 18-65% in the upland system. The implications of the results are discussed in terms of ino culation and management strategies.